EP3705610A1

EP3705610A1 - Carbon fiber bundle and method for producing same

Info

Publication number: EP3705610A1
Application number: EP18874618.4A
Authority: EP
Inventors: Daisuke Saito; Keiichi ISHIO; Koki IHARA
Original assignee: Toray Industries Inc
Current assignee: Toray Industries Inc
Priority date: 2017-10-31
Filing date: 2018-10-16
Publication date: 2020-09-09
Also published as: WO2019087766A1; JP6575696B1; KR20200040797A; CN111263834A; JPWO2019087766A1; TW201923183A; CN111263834B; EP3705610A4; KR102142368B1; US20200385892A1

Abstract

The purpose of the present invention is to provide: a carbon fiber bundle which exhibits excellent filament shape stability when forming a composite material, and from which it is possible to obtain a carbon fiber composite material which exhibits high tensile strength; and a method for producing the same. To achieve this purpose, the provided carbon fiber bundle exhibits a tensile elasticity as a resin-impregnated strand of 265-300 GPa, a tensile strength as a resin-impregnated strand of 6.0 GPa or more, a knot strength of 820 N/mm² or more, a filament number of 30,000 or more, and an average tear distance of 600-850 mm, wherein the rate of change in filament width is 8% or less when the carbon fiber bundle is unraveled under the conditions stipulated in the description, and there are four or fewer sections per 1,000 m which exhibit a filament width equal to or less than 75% of the average filament width when the carbon fiber bundle is unraveled under the conditions stipulated in the description.

Description

TECHNICAL FIELD

The present invention relates to a carbon fiber bundle that has high tensile strength and is suitable for being molded into a carbon fiber-reinforced composite (hereinafter sometimes simply referred to as a composite) in particular by a filament winding (hereinafter abbreviated as FW) molding method, and a method for producing the same.

BACKGROUND ART

Carbon fibers are characterized by excellent mechanical properties, in particular, high specific strength and high specific modulus. Therefore, carbon fibers are widely used in general industrial applications such as aerospace applications, leisure applications, and automobiles, and various methods for molding carbon fibers have been developed. Among the molding methods, the FW molding method has been widely applied to carbon fibers because of its excellent molding capability and the characteristics of a composite obtained from the carbon fibers. In particular, in order to obtain lightweight characteristics and high-performance characteristics in fuel containers of natural gas automobiles and the like that have been attracting attention in recent years, carbon fibers intended for reinforcing fibers and molded by the FW molding method have begun to be used. Furthermore, in recent years, there has been a growing demand for carbon fibers that are used at higher pressures than before and are suitable for FW molding applications, such as carbon fibers used in compressed hydrogen gas containers that are assumed to be filled with hydrogen gas for fuel cell applications.
For example, the working pressure of a compressed hydrogen gas container is 50 to 100 MPa, which is higher than that of a conventional compressed natural gas container of about 20 to 30 MPa. In particular, in a compressed hydrogen gas container for automobiles, it is desired to reduce the weight of the container in order to increase the mileage of automobiles. Therefore, attempts have been made to reduce the weight of the container with use of a high-strength carbon fiber-reinforced composite to reduce the amount of use of the carbon fiber-reinforced composite. Accordingly, carbon fiber-reinforced composites used in such applications are desired to have high strength and to be improved in stability as well as uniformity when molded by FW molding.
In general, in order for a composite to develop high tensile strength, it is important that the carbon fiber bundle have high tensile strength and high tensile modulus. For this reason, carbon fiber bundles having a number of filaments less than 30,000 are mainly produced for applications that emphasize excellent mechanical properties.
In brittle materials such as carbon fibers, it is possible to increase the tensile strength of a carbon fiber bundle by decreasing the flaw size of carbon fibers according to the Griffith equation or increasing the fracture toughness of carbon fibers. In particular, improvement of the fracture toughness of carbon fibers is effective in that the tensile strength of the carbon fiber bundle can be increased independent of the state of the flaw size of carbon fibers (Patent Document 1). Further, improvement of the fracture toughness of carbon fibers is effective also in that it is possible to efficiently increase the tensile strength of the carbon fiber-reinforced composite obtained using the carbon fibers, and to reduce fuzz that lowers the tensile strength of the composite.
Until now, as a method for improving the tensile strength and the tensile modulus of the carbon fiber bundle, there have been proposed a method of increasing the stabilization temperature using a plurality of ovens having different temperatures in the stabilization process, and a method of extending, in oxidation ovens composed of a plurality of ovens, precursor fibers for carbon fiber that have passed through the ovens according to the density thereof (Patent Documents 2 to 5). There has also been proposed a method of performing temperature control in two or three temperature control regions different in temperature in the stabilization process (Patent Document 6) .
In addition, a carbon fiber bundle having a large number of filaments and excellent in productivity has been proposed (Patent Documents 7 to 9).
Further, there has also been proposed a carbon fiber bundle having high knot strength, the carbon fiber bundle reflecting mechanical performance of the carbon fiber bundle in a direction other than the fiber axis direction, and exhibiting sufficient mechanical performance in a pseudoisotropic material (Patent Document 10).
As for the stability of tensile strength, there has been proposed a technique for improving the stability by selecting a specific copolymerization component for the precursor fiber for carbon fiber (Patent Document 11).
The FW molding method has originally been applied to glass fibers. It has been clarified that if a conventional carbon fiber bundle is used as it is in the FW molding method, fluctuation of the yarn shape of the strand, that is, fluctuation of the width of fiber bundle greatly affects the quality and composite properties of the molded product, because the number of filaments per carbon fiber strand is large. Meanwhile, a multifilament carbon fiber bundle having a large total fineness is desired for shortening the molding time for molding into a composite, and a fiber bundle having a stable width in an unwound state in spite of the large number of filaments has been proposed (Patent Document 9).

PRIOR ART DOCUMENTS

PATENT DOCUMENTS

Patent Document 1: International Publication No. 97/45576
Patent Document 2: Japanese Patent Laid-open Publication No. 58-163729
Patent Document 3: Japanese Patent Laid-open Publication No. 6-294020
Patent Document 4: Japanese Patent Laid-open Publication No. 62-257422
Patent Document 5: Japanese Patent Laid-open Publication No. 2013-23778
Patent Document 6: Japanese Patent Laid-open Publication No. 2012-82541
Patent Document 7: Japanese Patent Laid-open Publication No. 2005-113296
Patent Document 8: Japanese Patent Laid-open Publication No. 2005-60871
Patent Document 9: Japanese Patent Laid-open Publication No. 2012-154000
Patent Document 10: Japanese Patent Laid-open Publication No. 2015-96664
Patent Document 11: Japanese Patent Laid-open Publication No. 2015-71722

SUMMARY OF THE INVENTION

PROBLEMS TO BE SOLVED BY THE INVENTION

It is important to increase the fracture toughness of carbon fibers. To increase the fracture toughness, control of a microstructure of carbon fibers is essentially important. The proposal of Patent Document 1 is merely aimed at controlling the silicone oil agent, the single-fiber fineness, and the difference between skin-core structure, and improving the physical properties through the control of surface flaws or control of microstructure distribution of carbon fibers, and is not aimed at improving the microstructure itself.
In the proposal of Patent Document 2, the number of temperature control regions in the stabilization process is two or three, and the carbon fiber bundle is to be treated at a temperature as high as possible in the regions. The treatment time, however, is as long as 44 to 60 minutes, and the technique does not achieve the control of the microstructure region of the carbon fibers. In the proposal of Patent Document 3, the number of temperature control regions in the stabilization process is two or three, and the heat treatment time in the high temperature region is prolonged to achieve the stabilization in a short time. Therefore, the technique is inadequate in that the stabilization time at high temperature is long, and that the fiber structure at the initial stage of the stabilization is not controlled. The proposal of Patent Document 4 requires three to six ovens to set a plurality of degrees of extension in the oxidation ovens or to shorten the stabilization time, but does not achieve satisfactory control of the microstructure of carbon fibers. The proposal of Patent Document 5 is to set the specific gravity of fibers in the middle of the stabilization process to 1.27 or more and then heat-treat the carbon fiber bundle at 280 to 400°C for 10 to 120 seconds. The technique, however, does not achieve satisfactory control of the microstructure of carbon fibers merely by treating the carbon fiber bundle at a high temperature at the very last stage of the heat treatment. The proposal of Patent Document 6 is a technique of controlling the specific gravity of the stabilized yarn after the first oxidation oven to 1.27 or more, and does not achieve satisfactory control of the microstructure.
The proposal of Patent Document 7 is a technique in which a yarn is wet-spun from a spinneret having a large number of holes, and the stretch ratio in the spinning process is controlled. In the technique, however, the level of the tensile strength of resin-impregnated strands is low, and it is impossible to provide a composite that exhibits high tensile strength. Although the proposal of Patent Document 8 is a method of efficiently stabilizing a precursor fiber bundle for carbon fiber having a large number of filaments, the level of the tensile strength of resin-impregnated strands is low, and it is impossible to provide a composite that exhibits high tensile strength.
The proposal of Patent Document 9 is highly suitable for FW molding because of the stable width of fiber bundle in an unwound state in spite of the large number of filaments. The technique, however, does not achieve the control of the microstructure to control the fracture toughness of the carbon fiber bundle, and the document does not mention the knot strength and the coefficient of variation thereof.
Although the proposal of Patent Document 10 describes that the carbon fiber bundle has high knot strength mainly due to adjustment of the surface treatment of the carbon fiber bundle and the sizing agent, the document does not mention the number of filaments of the carbon fiber bundle, and the number of filaments is only 24,000 even in the examples. Since the knot strength decreases as the number of filaments of the carbon fiber bundle is increased in order to enhance the uniformity as the carbon fiber bundle, the technique is incapable of achieving both the number of filaments and the knot strength of the carbon fiber bundle.
Although the proposal of Patent Document 11 describes the reduction in variation of the tensile strength of resin-impregnated strands, the obtained composite has a low level of strength, and does not exhibit high tensile strength.
In order to solve the above-mentioned problems, an object of the present invention is to provide a carbon fiber bundle that is excellent in stability of the yarn shape when molded into a composite and is capable of providing a carbon fiber-reinforced composite having high tensile strength, and a method for producing the same.

SOLUTIONS TO THE PROBLEMS

The present inventors uniformized the heat treatment, improved the fracture toughness of the single fibers, and controlled the entanglement in the fiber bundle while increasing the number of filaments and significantly improving the production efficiency. As a result, they found a method of obtaining a carbon fiber bundle that is increased in the tensile strength and improved in the quality to a level that is not achieved with conventional carbon fiber bundles, and accomplished the present invention.
In order to achieve the above-mentioned object, the carbon fiber bundle according to the present invention has the following characteristics.
More specifically, the present invention provides a carbon fiber bundle having a tensile modulus of resin-impregnated strands of 265 to 300 GPa, a tensile strength of resin-impregnated strands of 6.0 GPa or more, a knot strength of 820 N/mm² or more, a number of filaments of 30,000 or more, and an average tearable length of 600 to 850 mm, the carbon fiber bundle having, when unwound under conditions described in the description, a variation of fiber bundle width of 8% or less, the carbon fiber bundle having, per 1,000 m, 4 or less sites that have a width of fiber bundle of 75% or less based on the average width of the carbon fiber bundle unwound under conditions described in the description.
Such a carbon fiber bundle is suitably obtained by a method including: a gathering process of gathering, using a roller just before gathering guide and a gathering guide, precursor fiber bundles for carbon fiber entering the gathering guide at a distance between the roller just before gathering guide and the gathering guide of 12 times or more the fiber bundle pitch between the precursor fiber bundles for carbon fiber to give a polyacrylonitrile precursor fiber bundle for carbon fiber having a number of filaments of 30,000 or more and an average tearable length of 400 to 800 mm; a first stabilization process of stabilizing the polyacrylonitrile precursor fiber bundle for carbon fiber obtained in the gathering process for 8 to 25 minutes until the ratio of the peak intensity at 1453 cm^-1 to the peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.98 to 1.10 to give a fiber bundle; a second stabilization process of stabilizing the fiber bundle obtained in the first stabilization process for 20 to 35 minutes until the ratio of the peak intensity at 1453 cm^-1 to the peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.60 to 0.65 and the ratio of the peak intensity at 1254 cm^-1 to the peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.50 to 0.65; a pre-carbonization process of pre-carbonizing the fiber bundle obtained in the second stabilization process in an inert gas having a maximum temperature of 500 to 1200°C at a stretch ratio of 1.00 to 1.10; and a carbonization process of carbonizing the fiber bundle obtained in the pre-carbonization process in an inert gas having a maximum temperature of 1000 to 2000°C.

EFFECTS OF THE INVENTION

According to the present invention, owing to the control of the entanglement, it is possible to provide a carbon fiber bundle that is excellent in stability of the yarn shape when molded into a composite in spite of the large number of filaments, and that can provide a high-quality carbon fiber-reinforced composite that exhibits high tensile strength.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a schematic view showing a method for measuring the average tearable length.
Fig. 2 is a schematic view showing a device for measuring a fiber form of an unwound carbon fiber bundle. The part upper than the broken line is a lateral view of the device, and the part lower than the broken line is a top view of the device.
Fig. 3 is a schematic view showing a gathering process performed with a gathering guide.

EMBODIMENTS OF THE INVENTION

In the carbon fiber bundle according to the present invention, the number of filaments is 30,000 or more, preferably 35,000 or more. In the production of a composite by FW, the productivity depends on the speed of fiber bundle and the number of filaments. Therefore, a large number of filaments enable efficient production of the composite. A number of filaments of 30,000 or more is satisfactory from the viewpoint of productivity. The upper limit of the number of filaments is not particularly limited, but the larger the number of filaments is, the more remarkable the yarn break is due to the heat generation of the yarn during the stabilization process. Therefore, the number of filaments is preferably 50,000 or less.
The carbon fiber bundle according to the present invention has a tensile modulus of resin-impregnated strands of 265 to 300 GPa, preferably 270 to 295 GPa, more preferably 275 to 290 GPa. In the present invention, the "tensile modulus of resin-impregnated strands" represents the tensile modulus in the tensile test of resin-impregnated strands. A tensile modulus of resin-impregnated strands of 265 to 300 GPa is preferable because the carbon fiber bundle is excellent in the balance between the tensile modulus of resin-impregnated strands and the tensile strength of resin-impregnated strands. In particular, a tensile modulus of resin-impregnated strands controlled to 275 to 290 GPa easily provides a carbon fiber bundle excellent in tensile strength of resin-impregnated strands. The tensile modulus of resin-impregnated strands can be determined by the method described in "Strand tensile test of carbon fiber bundle" described later. In the test, the range of strain is set to 0.1 to 0.6%. The tensile modulus of resin-impregnated strands of the carbon fiber bundle can be controlled mainly by applying tension to the fiber bundle or by changing the carbonization temperature in any of heat treatment processes in the production process of the carbon fiber bundle.
The carbon fiber bundle according to the present invention has a tensile strength of resin-impregnated strands of 6.0 GPa or more, preferably 6.2 GPa or more, more preferably 6.4 GPa or more. In the present invention, the "tensile strength of resin-impregnated strands" represents the tensile strength in the tensile test of resin-impregnated strands. When the tensile strength of resin-impregnated strands is 6.0 GPa or more, a composite produced from the carbon fiber bundle has a potential to exhibit satisfactory tensile strength. The tensile strength of resin-impregnated strands can be determined by the method described in "Strand tensile test of carbon fiber bundle" described later. The upper limit of the tensile strength of resin-impregnated strands is not particularly limited, but is usually about 7.0 GPa from the viewpoint of productivity.
Further, the coefficient of variation (%) of the tensile strength of resin-impregnated strands that is represented by the ratio of the standard deviation to the average of the tensile strength of resin-impregnated strands ([standard deviation]/[average]) is preferably 4% or less, more preferably 3.5%, still more preferably 2.5% or less. The lower limit of the coefficient of variation of the tensile strength of resin-impregnated strands is most preferably 0.0%, but is usually about 1.0%. In the production of a pressure vessel by FW molding, fracture starts from a portion where the tensile strength of resin-impregnated strands is the lowest among portions varied in the tensile strength of resin-impregnated strands. Therefore, if the coefficient of variation of the tensile strength of resin-impregnated strands is large, the amount of use of carbon fiber bundles has to be increased in consideration of the expected minimum tensile strength even if the average tensile strength of resin-impregnated strands is high. Therefore, the large amount of carbon fiber bundles lead to increase in the mass of a tank. Meanwhile, reducing the coefficient of variation of the tensile strength of resin-impregnated strands can reduce the amount of use of carbon fiber bundles and further reduce the weight of the composite. When the coefficient of variation of the tensile strength of resin-impregnated strands is 4% or less, in the production of a composite using carbon fiber bundles, a satisfactory composite with small variation of the tensile strength among portions can be obtained, and the amount of use of carbon fiber bundles can be reduced. The parameters relating to the tensile strength of resin-impregnated strands, that is, the tensile strength of resin-impregnated strands and the coefficient of variation of the tensile strength of resin-impregnated strands can be controlled using the method for producing a carbon fiber bundle according to the present invention described later.
Furthermore, the carbon fiber bundle according to the present invention has a fiber bundle tensile strength (also simply abbreviated as "knot strength") of 820 N/mm² or more, the fiber bundle tensile strength being obtained by forming a knot at the midpoint of the carbon fiber bundle and subjecting the carbon fiber bundle to a fiber bundle tensile test. The knot strength is preferably 850 N/mm² or more, more preferably 900 N/mm² or more. The upper limit of the knot strength is not particularly limited, but is usually about 1100 N/mm². The knot strength can be determined by the method described in "Knot strength of carbon fiber bundle" described later. The knot strength is an indicator that reflects the mechanical properties of the fiber bundle in a direction other than the fiber axis direction. During the production process of the composite, a bending direction is applied to the carbon fiber bundle. When the number of filaments is increased in order to efficiently produce a composite, fuzz is generated and it tends to be difficult to increase the speed of fiber bundle during the production of the composite. However, high knot strength enables production of a high-quality composite even under conditions where the speed of fiber bundle is high. When the knot strength is 820 N/mm² or more, it is possible to reduce fuzz due to abrasion with guide parts or rollers and to increase the speed of fiber bundle during the FW molding process. In order to increase the knot strength of the carbon fiber bundle, it is preferable to control the structural parameters particularly in the stabilization processes and the pre-carbonization process within preferable ranges in the method for producing a carbon fiber bundle according to the present invention described later.
In the carbon fiber bundle according to the present invention, the coefficient of variation (%) of the knot strength that is represented by the ratio of the standard deviation to the average of the knot strength ([standard deviation]/[average]) is preferably 5% or less, more preferably 4% or less, still more preferably 2% or less. The lower limit of the coefficient of variation of the knot strength is most preferably 0.0%, but is usually about 1.0%. In the FW molding process, when the coefficient of variation of the knot strength is high, the knot strength may be partially low and fuzz is likely to be generated at the portion where the variation of the knot strength is large, and it tends to be difficult to increase the speed of fiber bundle during the production of the composite. However, a low coefficient of variation of the knot strength can provide a high-quality composite. A coefficient of variation of the knot strength of 5% or less can sufficiently suppress fuzzing in a usual FW molding process. The lower limit of the coefficient of variation of the knot strength is not particularly limited, and a lower coefficient of variation is capable of more effectively suppressing fuzz and improving the production efficiency. However, since the effect of suppressing fuzz is saturated at a coefficient of variation of the knot strength of about 2%, generation of fuzz can be effectively suppressed by controlling the coefficient of variation of the knot strength to 2% or less. The coefficient of variation of the knot strength can be determined by the method described in "Knot strength of carbon fiber bundle" described later. The knot strength and the coefficient of variation of the knot strength can be controlled using the method for producing a carbon fiber bundle according to the present invention described later.
The carbon fiber bundle according to the present invention preferably has a product E × d/W of 13.0 GPa or more, wherein d/W is the ratio of the single-fiber diameter d (µm) to the loop diameter W (µm) just before loop fracture as evaluated by a single-fiber loop test, and E (GPa) is the tensile modulus of resin-impregnated strands. The product E × d/W is more preferably 13.3 GPa or more, still more preferably 13.5 GPa or more. The single-fiber loop test is a technique of investigating the relation between the strain given to a single fiber and a fracture behavior such as single fiber fracture and buckling by deforming the single fiber into a loop shape. When a single fiber is deformed into a loop shape, compressive strain is given to the inside of the single fiber, and tensile strain is given to the outside of the single fiber. Since compression buckling occurs before tensile fracture, the single-fiber loop test is conventionally often used as a test method for the single fiber compressive strength of a carbon fiber bundle. The single-fiber loop test, however, can be used to evaluate a value regarded as the intrinsic vending strength of a carbon fiber bundle since the test evaluates the fracture strain. That is, d/W is a value proportional to strain, and the product of the value of d/W and the tensile modulus of resin-impregnated strands, E (the details will be described later) is a value corresponding to the strength. Although the tensile strength of the composite is sometimes not increased even if merely the tensile strength of resin-impregnated strands of the carbon fiber bundle is increased, the tensile strength of the composite can be effectively increased by increasing the value of E × d/W. The upper limit of E × d/W is not particularly limited, and it is sufficient to set the upper limit of E × d/W to 19.0 GPa. In addition, the parameter can be controlled using the method for producing a carbon fiber bundle according to the present invention described later.
The carbon fiber bundle according to the present invention preferably has a Weibull shape parameter m in the Weibull plot of the value of E × d/W of 12 or more, wherein the value of E × d/W is determined for 20 single fibers. The Weibull shape parameter m is more preferably 15 or more, still more preferably 17 or more. The Weibull plot is a technique widely used for evaluating the strength distribution, and the Weibull shape parameter m tells the spread of the distribution. In the present invention, the single fibers are numbered as 1, ..., i, ..., and 20 in the order of the smallest value to the largest value of E × d/W, and the numbers are plotted in the Weibull plot with ln(-ln(1 - (i - 0.5)/20)) as the ordinate and ln(E × d/W) as the abscissa. Herein, ln means a natural logarithm. In the case where the plot is linearly approximated by the least squares method, the Weibull shape parameter m is obtained as the slope of the line. The larger the Weibull shape parameter m is, the narrower the strength distribution is, and the smaller the Weibull shape parameter m is, the wider the strength distribution is. In the case of a general carbon fiber bundle, the Weibull shape parameter m of the tensile strength evaluated by a single fiber tensile test often has a value around 5. It is understood that such value is derived from the large distribution of flaw sizes. Meanwhile, although the detailed reason is not necessarily clear, in the case of the carbon fiber bundle according to the present invention, the Weibull shape parameter m of E × d/W is significantly larger than the value around 5, and a Weibull shape parameter m of 12 or more often makes it possible to produce a composite having high tensile strength.
The carbon fiber bundle according to the present invention preferably has a product E × d/W of 13.0 GPa or more, and a Weibull shape parameter m in the Weibull plot of E × d/W of 12 or more, wherein d/W is the ratio of the single-fiber diameter d to the loop diameter W just before loop fracture as evaluated by a single-fiber loop test, and E is the tensile modulus of resin-impregnated strands. When the carbon fiber bundle simultaneously satisfies both of these conditions, a composite having particularly high tensile strength can be obtained.
In the present invention, the average tearable length of the carbon fiber bundle is 600 to 850 mm, preferably 650 to 850 mm, more preferably 700 to 850 mm. The average tearable length is an indicator showing the degree of entanglement in a certain fiber bundle. As the fiber bundle is strongly entangled uniformly, the average tearable length is shorter, and when there is no entanglement or the fiber bundle is entangled nonuniformly, the average tearable length is longer. In the case where the carbon fiber bundle is strongly entangled uniformly, it is possible to increase the strength of the carbon fiber bundle in a long gauge length on the order of several meters. Further, in the case where the carbon fiber bundle is strongly entangled, the width of fiber bundle of the running fibers is highly stable during the FW molding processing, and a molded product with stable quality and composite properties can be obtained. Therefore, when the average tearable length of the carbon fiber bundle is 850 mm or less, it is possible to transmit high tension sufficiently between the fibers, to enhance the fiber alignment in the carbon fiber bundle, to make the stress transfer in the composite obtained from the carbon fiber bundle more uniform, and to stabilize the width of fiber bundle of the running fibers during the FW molding processing. If the average tearable length of the carbon fiber bundle is less than 600 mm, stress concentration points are formed, and the tensile strength of a composite obtained from the carbon fiber bundle may be decreased. Any means can be adopted as a means for achieving such an entangled state of the carbon fiber bundle as long as the above-mentioned numerical range can be achieved. In particular, treatment by spraying a fluid onto the carbon fiber bundle is preferably used.
As for the carbon fiber bundle according to the present invention, the carbon fiber bundle defined as described above has, when unwound, a variation of fiber bundle width of 8% or less, and also has, per 1,000 m, 4 or less sites that have a width of fiber bundle of 75% or less based on the average width of the unwound carbon fiber bundle. If the variation of fiber bundle width is large, the carbon fiber bundles are unevenly distributed in the molded product, resulting in large variation of composite properties. In particular, there may be a case where satisfactory characteristics cannot be obtained in a portion where there are few fibers and, for example, a large amount of carbon fiber bundles are necessary for satisfying the required characteristics in terms of tensile strength, so that it is difficult to achieve weight reduction of the molded product. Suppressing the fluctuation of the width of fiber bundle in the unwound carbon fiber bundle and reducing the points where there are sites having an extremely small width of fiber bundle can provide a composite having stable composite properties. When the variation of fiber bundle width is 8% or less, satisfactory stability of the composite properties can be obtained. The variation of fiber bundle width is more preferably 6% or less, still more preferably 4% or less. Meanwhile, satisfactory stability of the composite properties can also be obtained when the carbon fiber bundle has, per 1,000 m, 4 or less sites that have a width of fiber bundle of 75% or less based on the average width of the unwound carbon fiber bundle. The number of sites is more preferably 3 or less, still more preferably 2 or less per 1,000 m.
Since the carbon fiber bundle according to the present invention is excellent in stability of the yarn shape when molded into a composite, the carbon fiber bundle is capable of providing a carbon fiber-reinforced composite having high tensile strength. Moreover, use of the carbon fiber bundle according to the present invention makes it easy to provide a carbon fiber-reinforced composite having high tensile strength and small variation of the tensile strength.
Then, a method for producing a carbon fiber bundle suitable for obtaining the carbon fiber bundle according to the present invention will be described.
It is preferable to use a polyacrylonitrile copolymer as a raw material used in the production of the precursor fiber bundle for carbon fiber. In the present invention, the "polyacrylonitrile copolymer" refers to a material containing at least acrylonitrile as a main component of a polymer unit. The "main component" usually refers to a component that accounts for 90 to 100 mass% of the polymer unit. In the production of the precursor fiber bundle for carbon fiber, the polyacrylonitrile copolymer preferably contains a copolymerization component from the viewpoint of controlling the stabilization treatment defined in the present invention.
A preferable example of a monomer usable as a copolymerization component is a monomer containing at least one carboxylic acid group or amide group from the viewpoint of accelerating the stabilization. Examples of the monomer containing a carboxylic acid group include acrylic acid, methacrylic acid, itaconic acid, and alkali metal salts and ammonium salts thereof. Examples of the monomer containing an amide group include acrylamide.
In the production of the precursor fiber bundle for carbon fiber, the method for producing the polyacrylonitrile copolymer can be selected from known polymerization methods.
In the production of the precursor fiber bundle for carbon fiber, either of a dry-jet wet spinning method and a wet spinning method may be used as the spinning method. A dry-jet wet spinning method that is advantageous in terms of the knot strength of the obtained carbon fiber bundle is preferably used. The spinning process includes: an extruding process of extruding a spinning dope solution from a spinneret into a coagulation bath and spinning the dope solution by the dry-jet wet spinning method to produce a fiber; a water washing process of washing the fiber obtained in the extruding process in a water bath; a water bath stretching process of stretching the fiber obtained in the water washing process in the water bath; and a drying heat treatment process of subjecting the fiber obtained in the water bath stretching process to drying heat treatment. The spinning process preferably also include, as necessary, a steam stretching process of steam-stretching the fiber obtained in the drying heat treatment process. The spinning process preferably also include, as necessary, a gathering process so that the number of filaments of the precursor fiber bundle for carbon fiber may be equal to the number of filaments of the carbon fiber bundle. Note that the order of these processes can be appropriately changed. The spinning dope solution is obtained by dissolving the above-mentioned polyacrylonitrile copolymer in a solvent capable of dissolving polyacrylonitrile, such as dimethylsulfoxide, dimethylformamide, and dimethylacetamide.
The coagulation bath preferably contains a solvent used as a solvent of the spinning dope solution, such as dimethylsulfoxide, dimethylformamide, and dimethylacetamide, and a coagulant. As the coagulant, those that do not dissolve the polyacrylonitrile copolymer and are compatible with the solvent used in the spinning dope solution can be used. Specifically, water is preferably used as the coagulant.
The water washing bath used in the water washing process is preferably a water washing bath having a temperature of 30 to 98°C and having a plurality of stages.
The stretch ratio in the water bath stretching process is preferably 2 to 6.
After the water bath stretching process, it is preferable to apply an oil agent made of silicone or the like to the fiber bundle for the purpose of preventing adhesion between the single fibers. The silicone oil agent is preferably modified silicone, and is preferably one containing highly heat-resistant amino-modified silicone.
The drying heat treatment process can be performed by a known method. For example, an example of the drying temperature is 100 to 200°C.
A precursor fiber bundle for carbon fiber suitable for providing the carbon fiber bundle according to the present invention can be obtained by steam-stretching the fiber as necessary after the water washing process, the water bath stretching process, the oil agent application process, and the drying heat treatment process. The steam stretching is preferably performed in pressurized steam at a stretch ratio of 2 to 6.
The number of filaments of the precursor fiber bundle for carbon fiber is preferably 30,000 or more, more preferably 35,000 or more so as to be equal to the number of filaments of the carbon fiber bundle. When the number of filaments of the precursor fiber bundle for carbon fiber is equal to the number of filaments of the carbon fiber bundle, voids between single fibers in the carbon fiber bundle, so-called bundle splitting, tend to be eliminated. The larger the number of filaments in the precursor fiber bundle for carbon fiber is, the easier to reduce the physical property variation of the carbon fiber bundle is. Therefore, for example, when a spinneret having a number of holes smaller than the number of filaments of the carbon fiber bundle, such as a spinneret having a number of holes of 300 to 15,000 is used, it is preferable to provide, in the precursor fiber bundle production process, the gathering process of equalizing the number of filaments of the precursor fiber bundle for carbon fiber to that of the carbon fiber bundle.
In order to obtain carbon fibers having a predetermined average tearable length, in the gathering process, fibers in precursor fiber bundles entering the gathering guide are preferably gathered using a roller just before gathering guide and a gathering guide at a distance between the roller just before gathering guide and the gathering guide of 12 times or more the fiber bundle pitch between the precursor fiber bundles for carbon fiber. The distance between the roller just before gathering guide and the gathering guide is more preferably 14 times or more the fiber bundle pitch between the precursor fiber bundles for carbon fiber.
It is also preferable to control the average tearable length of the precursor fiber bundle for carbon fiber to 400 to 800 mm. Controlling the average tearable length of the precursor fiber bundle for carbon fiber within the above-mentioned range makes it possible to uniformize the tension applied inside the fiber bundle during the production of the carbon fiber bundle among the fibers in the bundle and, for example, to maintain the change of the crystal orientation caused by the heat treatment uniform between the single fibers. Therefore, variation of quality in the carbon fiber bundle can be reduced. In order to control the average tearable length of the carbon fiber bundle, it is preferable to control the average tearable length of the precursor fiber bundle for carbon fiber. In order to reduce the unevenness of tension in the fiber bundle, an average tearable length of 800 mm or less is sufficient. A shorter average tearable length is preferable because the heat treatment of the fiber bundle can be performed uniformly. If the average tearable length is less than 400 mm, stress concentration points tend to be formed in the fiber bundle.
In order to control the average tearable length within the above-mentioned range, the gathering guide is required to be disposed at the above-mentioned position, and it is preferable to make the precursor fiber bundles for carbon fiber further pass through a second gathering process of spraying a fluid onto the precursor fiber bundles for carbon fiber.
Herein, the "gathering guide" used in the gathering process refers to a guide that includes a plurality of roller groups, and that stacks two or more precursor fiber bundles for carbon fiber on each other while rotating the precursor fiber bundles for carbon fiber by substantially 90°, and then contact the precursor fiber bundles for carbon fiber with the rollers a plurality of times so that the single fibers in the fiber bundles may be moved by the folding and twisting of the fiber bundles to be gathered into one precursor fiber bundle for carbon fiber. An example of the gathering guide is shown in Fig. 3. A roller 15 just before gathering guide refers to a gathering guide first roller 16 that is included in a roller group making up the gathering guide (the roller group includes the gathering guide first roller 17, a gathering guide second roller 18, and a gathering guide third roller 19) and with which precursor fiber bundles 17 for carbon fiber first come into contact, and a roller that is just before the gathering guide first roller 16 and in the precursor fiber bundle production process. The distance L between the roller 15 just before gathering guide and the gathering guide is the center-to-center distance between the roller 15 just before gathering guide and the gathering guide first roller 16. Further, the fiber bundle pitch Y between the precursor fiber bundles 17 for carbon fiber entering the gathering guide refers to the center-to-center distance between the precursor fiber bundles 17 for carbon fiber that are running adjacent to each other on the roller 15 just before gathering guide, and is a value determined by measuring the distance with a ruler.
The second gathering process refers to a process of performing a treatment to spray a fluid onto the precursor fiber bundles for carbon fiber. The fluid used in the second gathering process may be either of gas and liquid, but an inexpensive fluid such as air or nitrogen is preferable. In the treatment with a fluid, the fluid is preferably sprayed onto the fiber bundles using a nozzle. The shape of the nozzle for spraying the fluid is not particularly limited, but it is preferable to use a nozzle having 2 to 8 spouts. Although the arrangement of the spouts is not particularly limited, it is preferable to arrange an even number of spouts so as to surround the fiber bundles so that the angle formed by the longitudinal direction of the fiber bundles and the spraying direction of the fluid may fall within the range of 88° to 90°, and so that the spouts may form a pair or pairs facing each other. Other conditions such as the tension of the fiber bundles at the time of fluid spraying and the fluid extrusion pressure should be considered so as to adjust the average tearable length as appropriate.
In the process for producing a precursor fiber bundle for carbon fiber including the gathering process, when the second gathering process is provided in order to control the tearable length of the gathered precursor fiber bundle for carbon fiber, it is possible to provide the second gathering process before the gathering process having the gathering guide and perform the treatment of spraying the fluid onto the fiber bundles before entering the gathering guide, or to provide the second gathering process after the gathering process having the gathering guide and perform the treatment of spraying the fluid onto the gathered fiber bundle. Further, the second gathering process may be provided before and after the gathering process having the gathering guide.
The single-fiber fineness of the precursor fiber bundle for carbon fiber is preferably 0.5 to 1.5 dtex, more preferably 0.5 to 0.8 dtex from the viewpoint of increasing the tensile strength of resin-impregnated strands and the tensile modulus of resin-impregnated strands of the carbon fiber bundle. The coefficient of variation (%) of the basis weight of the polyacrylonitrile precursor fiber bundle for carbon fiber that is represented by the ratio of the standard deviation to the average of the basis weight ([standard deviation]/[average]) is preferably 1 to 4%. If the coefficient of variation is 4% or more, variation of the tensile strength of resin-impregnated strands and the tensile modulus of resin-impregnated strands is likely to be increased due to variation of the basis weight, and it is difficult to obtain a satisfactory composite.
In the method for producing a carbon fiber bundle, a carbon fiber bundle is obtained by subjecting a precursor fiber bundle for carbon fiber to stabilization processes, a pre-carbonization process, and a carbonization process. In order to increase the knot strength of the carbon fiber bundle and reduce the variation of the knot strength, at the time of subjecting the precursor fiber bundle for carbon fiber to the stabilization processes, the conditions are preferably controlled so that the obtained stabilized fiber may have a ratio of the peak intensity at 1453 cm^-1 to the peak intensity at 1370 cm^-1 in the infrared spectrum falling within the range of 0.60 to 0.65 and a ratio of the peak intensity at 1254 cm^-1 to the peak intensity at 1370 cm^-1 in the infrared spectrum falling within the range of 0.50 to 0.65. Peaks at 1453 cm^-1 in the infrared spectrum are derived from alkene, and decrease with the progress of stabilization. Peaks at 1370 cm^-1 and peaks at 1254 cm^-1 are peaks derived from stabilized structures (thought to be a naphthyridine ring structure and a hydrogenated naphthyridine ring structure, respectively), and increase with the progress of stabilization. In the stabilization processes, in general, peaks derived from polyacrylonitrile are decreased as much as possible to increase the carbonization yield. In the present invention, however, the conditions of the stabilization processes are set so as to intentionally leave many alkenes. A stabilized fiber bundle having such a structure is subjected to a pre-carbonization process to give the carbon fiber bundle according to the present invention. Further, it is important to set the stabilization conditions so that the ratio of the peak intensity at 1254 cm^-1 to the peak intensity at 1370 cm^-1 may fall within the range of 0.50 to 0.65. Peaks at 1254 cm^-1 are frequently observed at portions where the fiber bundle is insufficiently stabilized. When there are a large number of the structures, the knot strength tends to decrease. The peak intensity ratio decreases with the progress of stabilization, and the decrease at the initial stage is particularly large. Depending on the stabilization conditions, however, the peak intensity ratio may not fall within the range of 0.65 or less even if the time is increased.
In order to satisfy these two peak intensity ratios within the intended ranges, the conditions should be set with attention being mainly paid to that the amount of the copolymerization component contained in the polyacrylonitrile copolymer that constitutes the precursor fiber bundle for carbon fiber is small, that the precursor fiber bundle for carbon fiber has a small fineness, and that the stabilization temperature is increased at the latter stage. Specifically, it is preferable that the precursor fiber bundle for carbon fiber be heat-treated until the ratio of the peak intensity at 1453 cm^-1 to the peak intensity at 1370 cm^-1 in the infrared spectrum falls within the range of 0.98 to 1.10 (first stabilization process), and then heat-treated until the ratio of the peak intensity at 1453 cm^-1 to the peak intensity at 1370 cm^-1 in the infrared spectrum falls within the range of 0.60 to 0.65 and the ratio of the peak intensity at 1254 cm^-1 to the peak intensity at 1370 cm^-1 in the infrared spectrum falls within the range of 0.50 to 0.65 preferably at a temperature higher than that in the first stabilization process for a stabilization time of 20 to 35 minutes, preferably for 20 to 30 minutes (second stabilization process). In order to shorten the stabilization time in the second stabilization process, the stabilization temperature should be adjusted to a high temperature. An appropriate stabilization temperature depends on the characteristics of the precursor fiber bundle for carbon fiber. It is preferable to control the center temperature of the precursor fiber bundle for carbon fiber preferably to 250 to 300°C, more preferably to 250 to 280°C, still more preferably to 250 to 270°C to control the peak intensity ratios within the above-mentioned ranges of the infrared spectrum. The stabilization temperature does not have to be constant, and multistage temperature setting may be employed. In the case where there are three or more oxidation ovens, the treatment performed in the second and subsequent oxidation ovens is referred to as the second stabilization process. In the present invention, there is no limitation on the number of oxidation ovens to perform the second stabilization process. In order to increase the knot strength of the obtained carbon fiber bundle, it is preferable to increase the stabilization temperature and shorten the stabilization time. In the first stabilization process, it is preferable to perform the stabilization preferably for a stabilization time of 8 to 25 minutes, more preferably for 8 to 15 minutes at a stabilization temperature within the above-mentioned range.
The "stabilization time" as used herein means the time during which the fiber bundle stays in the oxidation oven, and the "stabilized fiber bundle" means a fiber bundle after the stabilization processes and before the pre-carbonization process. In addition, the "peak intensity" as used herein is the absorbance at each wavelength that is obtained by sampling a small amount of the stabilized fiber, measuring the infrared spectrum of the fiber, and subjecting the obtained infrared spectrum to baseline correction, and the spectrum is not subjected to peak splitting. Further, the sample for measurement is diluted with KBr so that the sample may have a concentration of 0.67 mass%. As described above, the conditions of stabilization should be considered according to the preferable production method described later by measuring the infrared spectrum every time the stabilization condition settings are changed. Appropriate control of the infrared spectrum peak intensity ratios of the stabilized fiber enables control of the knot strength of the obtained carbon fiber bundle.
In the present invention, the stabilization process means to heat-treat the precursor fiber bundle for carbon fiber at 200 to 300°C in an atmosphere containing oxygen.
The total treatment time of the stabilization processes can be appropriately selected preferably within the range of 28 to 55 minutes. More preferably, the total treatment time is selected within the range of 28 to 45 minutes.
In the pre-carbonization process of pre-carbonizing the fiber bundle obtained in the stabilization processes, the obtained stabilized fiber is preferably heat-treated in an inert gas at a maximum temperature of 500 to 1200°C. The stretch ratio in the pre-carbonization process is preferably 1.00 to 1.10, more preferably 1.03 to 1.07. In such a temperature range, the microstructure hardly suffers from flaws due to stretching. When the stretch ratio in the pre-carbonization process is 1.00 or more, the reaction of forming the initial carbonized structure between the molecules inside the fiber is promoted, and a dense fiber structure can be formed. As a result, the knot strength of the carbon fiber bundle can be increased. If the stretch ratio in the pre-carbonization process exceeds 1.10, high tension may be applied to the pre-carbonized fiber bundle to generate fuzz in some cases.
The specific gravity of the fiber bundle obtained through the pre-carbonization process is preferably 1.5 to 1.8.
The pre-carbonized fiber bundle is carbonized in an inert gas at a maximum temperature of 1000 to 2000°C. From the viewpoint of increasing the tensile modulus of resin-impregnated strands of the obtained carbon fiber bundle, it is preferable that the maximum temperature of the carbonization process be higher. However, too high a temperature may decrease the knot strength. Therefore, it is preferable to set the maximum temperature in consideration of both the conditions. The maximum temperature is more preferably 1200 to 1800°C, still more preferably 1200 to 1600°C.
The carbon fiber bundle obtained as described above is preferably subjected to oxidation treatment. The oxidation treatment introduces an oxygen-containing functional group into the carbon fiber bundle. The electrochemical treatment of fiber surface employed in the present invention may be vapor phase oxidation, liquid phase oxidation, or liquid phase electrolytic oxidation. Liquid phase electrolytic oxidation is preferably employed from the viewpoint of high productivity and the capability of uniform treatment. In the present invention, the method of liquid phase electrolytic oxidation is not particularly limited, and a known method may be employed.
After the electrolytic treatment, the obtained carbon fiber bundle can be subjected to sizing treatment for imparting convergency to the carbon fiber bundle. For the sizing agent, a sizing agent well compatible with the matrix resin used in the composite can be appropriately selected according to the type of the matrix resin.
Methods for measuring various physical properties used in the present invention are as follows.

A package 7 of carbon fiber bundles is supplied to a creel 8 having a yarn path shown in Fig. 2, and the carbon fiber bundle is guided to a yarn path F shown in the drawing. In order to stabilize the yarn path, the yarn path F has a structure to twist the carbon fiber bundle once by 90° with a yarn path regulating roller 9, then twist the carbon fiber bundle in the opposite direction and pass the carbon fiber bundle through a flat roller group 10, then pass the carbon fiber bundle through an optical sensor 11 for measuring the width of fiber bundle, then pass the carbon fiber bundle through a drive roller group 13 for controlling the speed of fiber bundle, and wind the carbon fiber bundle on a winder 14. After the carbon fiber bundle is guided to the yarn path, the package 7 of carbon fiber bundles is unwound with the conditions being set to predetermined conditions, that is, the tension for drawing the carbon fiber bundle from the creel being set to 6 N/tex and the speed of fiber bundle passing through the yarn path being set to 50 m/min or more. During unwinding under the predetermined conditions, the width of the carbon fiber bundle in the air is measured with the optical sensor 11, and the average width of the carbon fiber bundle and the variation of fiber bundle width calculated from the ratio of the average to the standard deviation are determined as to the carbon fiber bundle unwound over 1,000 m or more. Moreover, for the carbon fiber bundle unwound over 1,000 m or more, the number of sites where the width of fiber bundle is 75% or less based on the average width of the unwound carbon fiber bundle are counted, and it is determined as to whether the number of sites where the width of fiber bundle is 75% or less based on the average width of the unwound carbon fiber bundle is 4 or less per 1,000 m. Herein, data detected by the optical sensor 11 are taken at intervals of 0.1 seconds using a data logger of NR600 or NR1000 manufactured by KEYENCE CORPORATION.

<Single-fiber loop test>

A single fiber having a length of about 10 cm is placed on a slide glass, 1 to 2 drops of glycerin is dropped on the center of the single fiber, and both ends of the single fiber are lightly twisted in the circumferential direction of the fiber to form a loop at the center of the single fiber. A cover glass is placed on the single fiber. The obtained specimen is put on a stage of a microscope, and shooting of a moving image is started under the conditions of a total magnification of 100 times and a frame rate of 15 frames/second. While adjusting the stage as appropriate so that the loop may not come out of the field of view, strain is applied to the single fiber until the single fiber fractures by pulling both the ends of the looped fiber at a constant speed in opposite directions with the ends being pushed against the slide glass with fingers. The frame just before loop fracture is specified by frame advance, and the width W of the loop just before loop fracture is measured by image analysis. The fiber diameter d is divided by W to calculate d/W. The number of tests n is 20. The value of E × d/W is obtained by multiplying the average of d/W by the tensile modulus of resin-impregnated strands.

The tensile modulus of resin-impregnated strands, E and the tensile strength of resin-impregnated strands of the carbon fiber bundle are determined according to the "resin-impregnated strand test method" of JIS R7608 (2008). The tensile modulus E of resin-impregnated strands is measured under a strain in the range of 0.1 to 0.6%. The test piece is produced by impregnating the carbon fiber bundle with the following resin composition, and under the curing conditions of heat treatment at a temperature of 130°C for 35 minutes.

[Resin composition]

3,4-Epoxycyclohexylmethyl-3,4-epoxy-cyclohexanecarboxylate (100 parts by mass)
Boron trifluoride monoethylamine (3 parts by mass)
Acetone (4 parts by mass)

The number of strands measured is 10, and the arithmetic averages of the measurement results are regarded as the tensile modulus of resin-impregnated strands and the tensile strength of resin-impregnated strands of the carbon fiber bundle. Further, as for the tensile strength, the standard deviation of 10 strands is obtained and divided by the average, and then the coefficient of variation is calculated as a percentage ([standard deviation]/[average] × 100). In the examples and comparative examples described later, "BAKELITE (registered trademark)" ERL-4221 manufactured by Union Carbide Corporation is used as the 3,4-epoxycyclohexylmethyl-3,4-epoxy-cyclohexane-carboxylate. The strain is measured using an extensometer.

A grip having a length of 25 mm is attached to both ends of a carbon fiber bundle having a length of 150 mm to produce a test specimen. In the production of the test specimen, a load of 0.1 × 10^-3 N/denier is applied to the carbon fiber bundle for alignment. One knot is made at the midpoint of the test specimen, and the test specimen is subjected to a fiber bundle tensile test at a crosshead speed at tension of 100 mm/min. A total of 12 fiber bundles are subjected to the measurement. The average of 10 fiber bundles excluding the maximum value and the minimum value is used as the measured value, and the standard deviation of 10 values is used as the standard deviation of the knot strength. As the knot strength, a value obtained by dividing the maximum load value obtained in the tensile test by the average cross-sectional area of the carbon fiber bundles is used. As for the coefficient of variation of the knot strength, the ratio between the knot strength of the carbon fiber bundles described above and the standard deviation of the knot strength is obtained, and a value expressed as a percentage is used ([standard deviation]/[average] × 100).

A stabilized fiber to be measured is frozen and pulverized, and then 2 mg of the stabilized fiber is accurately weighed and collected. The stabilized fiber is well mixed with 300 mg of KBr, and the mixture is placed in a molding jig and pressurized with a pressing machine at 40 MPa for 2 minutes to produce a tablet for measurement. The tablet is set in a Fourier transform infrared spectrophotometer, and the spectrum of the tablet is measured in the range of 1000 to 2000 cm^-1. The background correction is performed by subtracting from each intensity the minimum value thereof so that the minimum value in the range of 1700 to 2000 cm^-1 may be zero. The spectrophotometer used as the Fourier transform infrared spectrophotometer is Paragon 1000 manufactured by PerkinElmer Japan Co., Ltd.

The average tearable lengths of the precursor fiber bundle for carbon fiber and the carbon fiber bundle are both determined as follows. That is, as shown in Fig. 1, a fiber bundle 1 to be measured is cut into a length of 1160 mm, and one end 2 of the fiber bundle 1 is fixed to a horizontal table with an adhesive tape (the point is called a fixed point A). One end 3 of the fiber bundle 1 that is not fixed is divided into two by finger, and one of the divided ends of the fiber bundle is fixed to the table with an adhesive tape so as not to move in a state where the fiber bundle is strained (the point is called a fixed point B). The other one of the end 3 of the divided fiber bundle is moved along the table with the fixed point A as a supporting point so as not to slack, stopped at a position 4 where the direct distance from the fixed point B is 500 mm, and fixed to the table with an adhesive tape so as not to move (the point is called a fixed point C). The region surrounded by the fixed points A, B, and C is visually observed, and an entanglement point 5 farthest from the fixed point A is found. The length obtained by projecting the entanglement point 5 on the straight line connecting the fixed points A and B is read using a ruler with a smallest scale of 1 mm as a tearable length 6. The measurement is repeated 30 times, and the arithmetic average of the measured values is taken as the average tearable length. In this measurement method, the entanglement point farthest from the fixed point A is a point that is farthest in direct distance from the fixed point A and where three or more single fibers are entangled with each other with no slack.

Against a fixed chromium-plated stainless steel rod having a diameter of 12 mm, 200 mm of a carbon fiber bundle is abraded in a direction perpendicular to the axial direction of the stainless steel rod from one end of the fiber bundle to the other end thereof with 500 gf of tension being applied to the carbon fiber bundle. In the abrasion, the carbon fiber bundle is abraded over a distance of half the circumference of the stainless steel rod. After the carbon fiber bundle is reciprocated 20 times and abraded against the stainless steel rod a total of 40 times, the abraded carbon fiber bundle is sandwiched between two urethane sponges. A weight of 125 g is put on the urethane sponges so that the load may be applied to the entire surface of the urethane sponges, and the mass of the fuzz attached to the sponges after the abraded carbon fiber bundle is passed at a speed of 2 m/min is evaluated as the amount of abrasive fuzz.

The strand tensile test described above is performed with the resin composition being changed as follows.

[Resin composition]

Resorcinol epoxy (100 parts by mass)
Diethylenetriamine (39 parts by mass)

The curing conditions are 100°C for 2 hours. For the measurement, the carbon fiber bundle abraded against the stainless steel rod in the measurement of the amount of fuzz is used. As the resorcinol epoxy, Denacol EX201 manufactured by Nagase ChemteX Corporation is used. As the diethylenetriamine, the one manufactured by Tokyo Chemical Industry Co., Ltd. is used.

EXAMPLES

(Example 1)

A copolymer consisting of 99.0 mass% of acrylonitrile and 1.0 mass% of itaconic acid was polymerized by solution polymerization using dimethylsulfoxide as a solvent to prepare a spinning dope solution containing a polyacrylonitrile copolymer. Coagulated fibers were obtained by a dry-jet wet spinning method of extruding the obtained spinning dope solution once into the air from a spinneret having 12,000 holes, and introducing the extruded spinning dope solution into a coagulation bath made of an aqueous solution of dimethylsulfoxide.
The coagulated fibers were washed with water by a common method and stretched in a water bath at a stretch ratio of 3.5. Then, to the fiber bundle obtained after the water bath stretching, an amino-modified silicone oil agent was applied, and the fiber bundle was subjected to drying densification treatment using a heating roller at 160°C. Then, the fiber bundle was stretched 3.7 times in pressurized steam to achieve a total stretch ratio in the spinning of 13. Then, the filaments were passed through a gathering guide disposed so that the distance between the roller just before gathering guide and the gathering guide might be 16 times the fiber bundle pitch between the fiber bundles entering the gathering guide to gather the filaments, whereby a precursor fiber bundle for carbon fiber having a number of single fibers of 36,000 was produced. The precursor fiber bundle for carbon fiber had a single-fiber fineness of 0.8 dtex, and a percentage of the coefficient of variation of basis weight ([standard deviation]/[average]) of 3%.
Then, the precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 240°C and a stabilization time of 17 minutes for the first stabilization process and a stabilization temperature of 269°C and a stabilization time of 28 minutes for the second stabilization process to produce a stabilized fiber bundle shown in Table 1.
The obtained stabilized fiber bundle was pre-carbonized in a nitrogen atmosphere having a maximum temperature of 900°C while being stretched at a stretch ratio shown in Table 1 to produce a pre-carbonized fiber bundle. The obtained pre-carbonized fiber bundle was carbonized in a nitrogen atmosphere at a maximum temperature of 1500°C while being stretched at a stretch ratio shown in Table 1. The obtained carbon fiber bundle was subjected to surface treatment and sizing agent coating treatment to finally obtain a carbon fiber bundle having an average tearable length of 742 mm, a variation of fiber bundle width when unwound under the above-mentioned conditions of 6.8%, and a number of sites where the width of fiber bundle is 75% or less based on the average width of the carbon fiber bundle unwound under the above-mentioned conditions of 0.5 per 1,000 m. The physical properties are shown in Table 1.

(Example 2)

A stabilized fiber bundle was obtained as in Example 1 by changing only the position of the gathering guide so that the distance between the roller just before gathering guide and the gathering guide might be 12 times the fiber bundle pitch between the fiber bundles entering the gathering guide, subjecting, after passage through the gathering guide, the fiber bundle to second gathering with the air at a fluid extrusion pressure of 0.29 MPa-G with a tension of 2 mN/dtex being applied to the fiber bundle to produce a precursor fiber bundle for carbon fiber having a number of filaments of 36,000, and further changing the stabilization processes as follows. The precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 244°C and a stabilization time of 20 minutes for the first stabilization process and a stabilization temperature of 270°C and a stabilization time of 23 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 except that the stretch ratio in the pre-carbonization was 1.06 to produce a carbon fiber bundle.

(Example 3)

A precursor fiber bundle for carbon fiber obtained in the same manner as in Example 1 was used. The precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 244°C and a stabilization time of 20 minutes for the first stabilization process and a stabilization temperature of 270°C and a stabilization time of 23 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 to produce a carbon fiber bundle. The obtained carbon fiber-reinforced composite had a tensile strength at 0° of 5.3 GPa.

(Reference Example 1)

A carbon fiber bundle was obtained in the same manner as in Example 2 except that the position of the gathering guide was changed so that the distance between the roller just before gathering guide and the gathering guide might be 20 times the fiber bundle pitch between the fiber bundles entering the gathering guide.

(Example 4)

A precursor stabilized yarn for carbon fiber having a number of filaments of 36,000 was obtained as in Example 1 by changing only the position of the gathering guide so that the distance between the roller just before gathering guide and the gathering guide might be 12 times the fiber bundle pitch between the fiber bundles entering the gathering guide. Using the precursor stabilized yarn for carbon fiber, the precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 240°C and a stabilization time of 20 minutes for the first stabilization process and a stabilization temperature of 275°C and a stabilization time of 23 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 to produce a carbon fiber bundle.

(Comparative Example 1)

A carbon fiber bundle was obtained in the same manner as in Example 1 except that only the position of the gathering guide was changed so that the distance between the roller just before gathering guide and the gathering guide might be 10 times the fiber bundle pitch between the fiber bundles entering the gathering guide in Example 1 to produce a precursor fiber bundle for carbon fiber having a number of filaments of 36,000.

(Comparative Example 2)

The results of evaluating the carbon fiber bundle Panex 35 (manufactured by ZOLTEK Corporation) are shown in Table 1.

(Comparative Example 3)

A precursor fiber bundle for carbon fiber having a number of filaments of 24,000 was obtained as in Example 1 by changing only the position of the gathering guide so that the distance between the roller just before gathering guide and the gathering guide might be 11 times the fiber bundle pitch between the fiber bundles entering the gathering guide, and a stabilized fiber bundle was obtained as in Example 1 by changing the stabilization processes as follows. The precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 240°C and a stabilization time of 36 minutes for the first stabilization process and a stabilization temperature of 250°C and a stabilization time of 37 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 to produce a carbon fiber bundle. The results of evaluating the carbon fiber bundle are shown in Table 1.

(Comparative Example 4)

The stabilization, pre-carbonization, and carbonization treatment were performed in the same manner as in Comparative Example 3 except that the number of filaments of the precursor fiber bundle for carbon fiber was adjusted to 12,000 in Comparative Example 3 to produce a carbon fiber bundle. The results of evaluating the obtained carbon fiber bundle are shown in Table 1.

(Comparative Example 5)

Two carbon fiber bundles of Comparative Example 4 each having a number of filaments of 12,000 were gathered, and the gathered bundle having a number of filaments of 24,000 was evaluated. The results are shown in Table 1.

(Comparative Example 6)

Three carbon fiber bundles of Comparative Example 4 each having a number of filaments of 12,000 were gathered, and the gathered bundle having a number of filaments of 36,000 was evaluated. The results are shown in Table 1. The carbon fiber-reinforced composite had a tensile strength at 0° of 5.0 GPa, which was lower than that of Example 3 having a comparable tensile strength of resin-impregnated strands.

(Comparative Example 7)

A stabilized fiber bundle was obtained as in Example 1 except that only the stabilization processes were changed as follows. The precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 245°C and a stabilization time of 15 minutes for the first stabilization process and a stabilization temperature of 255°C and a stabilization time of 44 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 to produce a carbon fiber bundle. The amount of abrasive fuzz of the obtained carbon fiber bundle was larger than those of the carbon fiber bundles mentioned in the examples, and the carbon fiber bundle did not exhibit carbonization characteristics at a sufficiently high level and had a tensile strength of resin-impregnated strands of 5.9 GPa and a knot strength of 785 N/mm².

(Comparative Example 8)

A stabilized fiber bundle was obtained as in Example 1 except that only the stabilization processes were changed as follows. The precursor fiber bundle for carbon fiber was subjected to stabilization treatment while being stretched at a stretch ratio of 1 in an oven in an air atmosphere under the conditions of a stabilization temperature of 230°C and a stabilization time of 36 minutes for the first stabilization process and a stabilization temperature of 245°C and a stabilization time of 71 minutes for the second stabilization process to produce a stabilized fiber bundle. The subsequent pre-carbonization treatment and carbonization treatment were performed in the same manner as in Example 1 to produce a carbon fiber bundle. The amount of abrasive fuzz of the obtained carbon fiber bundle was larger than those of the carbon fiber bundles mentioned in the examples, and the carbon fiber bundle did not exhibit carbonization characteristics at a sufficiently high level and had a tensile strength of resin-impregnated strands of 5.9 GPa and a knot strength of 814 N/mm².

(Comparative Example 9)

A carbon fiber bundle was obtained in the same manner as in Comparative Example 8 except that only the position of the gathering guide was changed so that the distance between the roller just before gathering guide and the gathering guide might be 16 times the fiber bundle pitch between the fiber bundles entering the gathering guide in Comparative Example 8.

[Table 1-1]

	After first stabilization	After second stabilization		First stabilization treatment time	Second stabilization treatment time	Number of filaments
	1453 cm^-1/ 1370 cm^-1	1453 cm^-1/ 1370 cm^-1	1254 cm^-1/ 1370 cm^-1	min	min	number
Example 1	1.01	0.64	0.59	17	28	36000
Example 2	0.99	0.64	0.60	20	23	36000
Example 3	0.99	0.64	0.60	20	23	36000
Reference Example 1	0.99	0.64	0.60	20	23	36000
Example 4	1.03	0.62	0.58	20	23	36000
Comparative Example 1	1.01	0.64	0.59	17	28	36000
Comparative Example 2	-	-	-	-	-	50000
Comparative Example 3	0.87	0.63	0.60	36	37	24000
Comparative Example 4	0.87	0.63	0.60	36	37	12000
Comparative Example 5	-	-	-	-	-	24000
Comparative Example 6	-	-	-	-	-	36000
Comparative Example 7	1.01	0.67	0.61	15	44	36000
Comparative Example 8	0.95	0.62	0.59	36	71	36000
Comparative Example 9	0.95	0.62	0.59	36	71	36000

[Table 1-2]

	Pre-carbonization stretch ratio	Carbonization stretch ratio	Tensile strength of resin-impregnated strands	Coefficient of variation of tensile strength of resin-impregnated strands	Tensile modulus of resin-impregnated strands	Knot strength	Coefficient of variation of knot strength
	-	-	GPa	%	GPa	N/mm²	%
Example 1	1.03	0.95	6.3	3.6	270	935	4
Example 2	1.06	0.95	6.3	3.2	266	940	5
Example 3	1.06	0.95	6.4	3.6	269	940	5
Reference Example 1	1.06	0.95	6.2	2.9	270	913	3
Example 4	1.03	0.95	6.2	5.5	268	901	5
Comparative Example 1	1.06	0.95	6.2	3.6	273	929	4
Comparative Example 2	-	-	4.1	4.5	237	340	14
Comparative Example 3	0.98	0.95	5.9	4.5	277	875	7
Comparative Example 4	0.98	0.95	6.2	2.9	290	836	10
Comparative Example 5	-	-	6.2	2.9	290	730	14
Comparative Example 6	-	-	6.2	2.9	290	792	16
Comparative Example 7	1.03	0.95	5.9	5.2	268	785	7
Comparative Example 8	1.03	0.95	5.9	4.8	270	814	12
Comparative Example 9	1.03	0.95	5.9	3.6	270	818	5

In the table, "Pre-carbonization stretch ratio" and "Carbonization stretch ratio" mean the stretch ratio in the pre-carbonization process and the stretch ratio in the carbonization process, respectively.

[Table 2]

	E × d/W	Weibull shape parameter m of E × d/W	Avexage tearable length	Variation of fiber bundle width	Number of sites where width of fiber bundle is 75% or less based on average width of carbon fiber bundle	Single-fiber diameter	Amount of fuzz	Position of gathering guide	Second gathering process	TPP processability
	GPa		mm	%	sites/1,000 m	µm	mg/200 mm	Position of gathering guide	Second gathering process	TPP processability
Example 1	13.9	17	742	6.8	0.5	5.7	0.32	16	No	Good
Example 2	138	17	639	3.7	0.6	56	0.60	12	Yes	Good
Example 3	138	17	703	54	2.5	5.6	0.50	16	No	Good
Reference Example 1	13.8	17	580	3.2	1.8	5.7	0.65	20	Yes	Good
Example 4	13.6	16	837	7 .7	3.5	5.7	0.92	12	No	Good
Comparative Example 1	13.9	17	871	8.7	14	56	0.65	10	No	Poor
Comparative Example 2	8.3	7	885	6.5	5.9	7.2	2.25	-	-	Poor
Comparative Example 3	12.3	14	841	5.5	3.9	5.5	0.75	11	No	Poor
Comparative Example 4	14.2	13	848	3.2	1.5	5.5	0.95	-	-	-
Comparative Example 5	-	-	848	-	-	-	1.11	-	-	-
Comparative Example 6	-	-	848	-	-	-	1.36	-	-	-
Comparative Example 7	12.4	11	983	10	18	5.6	1.25	8	No	Poor
Comparative Example 8	13.7	15	851	8.1	10	5.7	1.12	11	No	Poor
Comparative Example 9	13.7	15	750	3.8	4	5.7	1.12	16	No	Good

DESCRIPTION OF REFERENCE SIGNS

1: Fiber bundle
2: Fixed point A
3: Fixed point B
4: Fixed point C
5: Entanglement point
6: Tearable length
7: Package of carbon fiber bundles
8: Creel
9: Yarn path regulating roller
10: Flat roller group
11: Optical sensor
12: Flat roller
13: Drive roller
14: Winder
F: Yarn path through which carbon fiber bundle passes
15: Roller just before gathering guide
16: Gathering guide first roller
17: Precursor fiber bundle for carbon fiber
18: Gathering guide second roller
19: Gathering guide third roller
20: Space for fixing gathering guide rollers
L: Distance between roller just before gathering guide and gathering guide
Y: Fiber bundle pitch between precursor fiber bundles for carbon fiber before gathering

Claims

A carbon fiber bundle having a tensile modulus of resin-impregnated strands of 265 to 300 GPa, a tensile strength of resin-impregnated strands of 6.0 GPa or more, a knot strength of 820 N/mm² or more, a number of filaments of 30,000 or more, and an average tearable length of 600 to 850 mm,
the carbon fiber bundle having, when unwound under conditions described in the description, a variation of fiber bundle width of 8% or less,
the carbon fiber bundle having, per 1,000 m, 4 or less sites that have a width of fiber bundle of 75% or less based on an average width of the carbon fiber bundle unwound under conditions described in the description.
The carbon fiber bundle according to claim 1, having a product E × d/W of 13.0 GPa or more, and a Weibull shape parameter m in a Weibull plot of E × d/W of 12 or more, wherein d/W is a ratio of a single-fiber diameter d to a loop diameter W just before loop fracture as evaluated by a single-fiber loop test, and E is the tensile modulus of resin-impregnated strands.
The carbon fiber bundle according to claim 1 or 2, having a coefficient of variation of the knot strength of 5% or less, the coefficient of variation being represented by a ratio of a standard deviation to an average of the knot strength.
The carbon fiber bundle according to claim 3, having a coefficient of variation of the tensile strength of resin-impregnated strands of 4% or less, the coefficient of variation being represented by a ratio of a standard deviation to an average of the tensile strength of resin-impregnated strands.
A method for producing the carbon fiber bundle according to any one of claims 1 to 4, the method comprising:
a gathering process of gathering, using a roller just before gathering guide and a gathering guide, precursor fiber bundles for carbon fiber entering the gathering guide at a distance between the roller just before gathering guide and the gathering guide of 12 times or more a fiber bundle pitch between the precursor fiber bundles for carbon fiber to give a polyacrylonitrile precursor fiber bundle for carbon fiber having a number of filaments of 30,000 or more and an average tearable length of 400 to 800 mm;

a first stabilization process of stabilizing the polyacrylonitrile precursor fiber bundle for carbon fiber obtained in the gathering process for 8 to 25 minutes until a ratio of a peak intensity at 1453 cm^-1 to a peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.98 to 1.10 to give a fiber bundle;

a second stabilization process of stabilizing the fiber bundle obtained in the first stabilization process for 20 to 35 minutes until a ratio of a peak intensity at 1453 cm^-1 to a peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.60 to 0.65 and a ratio of a peak intensity at 1254 cm^-1 to a peak intensity at 1370 cm^-1 in an infrared spectrum falls within a range of 0.50 to 0.65;

a pre-carbonization process of pre-carbonizing the fiber bundle obtained in the second stabilization process in an inert gas having a maximum temperature of 500 to 1200°C at a stretch ratio of 1.00 to 1.10; and

a carbonization process of carbonizing the fiber bundle obtained in the pre-carbonization process in an inert gas having a maximum temperature of 1000 to 2000°C.
The method according to claim 5, comprising a second gathering process of spraying a fluid onto the precursor fiber bundles for carbon fiber before and/or after the gathering process.
The method according to claim 5 or 6, wherein the polyacrylonitrile precursor fiber bundle for carbon fiber has a coefficient of variation of a basis weight of 1 to 4%, the coefficient of variation being represented by a ratio of a standard deviation to an average of the basis weight.